The present invention relates generally to indexing of media files and more specifically to a technique for automatically associating index information with a media file.
With the growth of consumer electronic devices, specifically electronic and digital devices with recording or capture capabilities, there has been a corresponding increase in the number of captured media files. As the number of captured media files increases, it become increasingly difficult to remember the content of each file, much less organize or search such files based on an understanding of the content of each file. Associating index information with each file can facilitate such activities.
However, there currently does not exist a cost-effective and time efficient manner for indexing multiple media files. One current available method is for an individual to create a separate indexing system by manually applying indexing terms to the media files. If the indexing terms are maintained in a searchable format, the user may then execute searches for pertinent media files. This can be an extremely time-consuming task, i.e. for an individual to view, label and index all previously existing media files and also view, label and index all future media files when they are provided to a secondary storage device. Furthermore, because individual users may create their own indexing schemes, complications may arise when users share or exchange various media files.
Another currently available indexing system is for a third party to index a user's media files. Similar to the user-created index system, this process is extremely inefficient and time consuming and is therefore a costly endeavor. Moreover, this process requires a third party to inspect and catalogue all of a user's media files, which may be undesirable and expensive to the user.
The capture devices, such as digital cameras and camcorders, for example, may provide some relevant indexing information, but this information tends to be limited to such data as the date of capture and the device settings used. They may also allow the user to manually enter a small amount of additional information. Manual entry of the information may be cumbersome and confusing to the typical user. Also, this can be a time-consuming task, which provides limited benefit, unless the user provides these details for all media files.
Another approach to indexing media files employs machine vision techniques to recognize certain elements within a photograph or video. Machine visions approaches, however, currently are limited in their recognition ability. Even as machine vision techniques improve, it remains difficult for machine vision approaches to recognize the context of the subjects within the media file. For example, even if a machine vision approach could recognize that a photograph depicts a baseball game, it would not be possible to recognize that it a specific game, such as a critical play-off game.
Therefore, a need exists for a media file indexing system whereupon a media file may be automatically indexed, which index is searchable to retrieve pertinent media files.